Despite their widespread adoption and advantages, LMS platforms present various challenges, particularly related to UX. One primary issue is the rigidity of LMS platforms as they offer limited flexibility, which is needed for dynamic teaching and learning approaches. Chen (2019) criticises LMS platforms for being overly administrative and not sufficiently fostering student-centred, interactive learning experiences. Another significant challenge is the steep learning curve associated with LMS platforms. This necessitates effective and continuous training for users to be able to use these systems optimally (Allam et al. 2024; Almusharraf 2024). Inadequate maintenance and technical support can lead to substantial expenses and can potentially negatively impact user satisfaction and engagement (Lamo, Perales & De-la-Fuente 2022). Moreover, many LMS platforms offer limited support for mobile devices, which restricts accessibility and usability in an increasingly mobile-dependent world. Improving mobile-friendly interfaces is thus essential to supporting the growing use of smartphones and tablets for educational purposes (Az-Zahra, Nurhayati & Herlambang 2023).

The integration of UX principles into educational technology is important for creating engaging and effective LMSs (Maslov, Nikou & Hansen 2021). As the field of UX design continues to evolve, it encompasses a broad range of disciplines and methodologies aimed at enhancing the overall interaction between users and digital products. In the context of LMSs, understanding and implementing good UX practices can significantly impact user satisfaction and educational outcomes.

To ensure the cohesive integration of LMS and UX elements and considerations, it is essential to recognise that the effectiveness of LMS platforms is intrinsically linked to their usability and user experience. While LMS platforms provide the infrastructure for educational content and administrative functions, the success of these systems hinges on their ability to offer an intuitive, user-friendly experience that meets the diverse needs of users.

User satisfaction and engagement in LMS environments are influenced by several interconnected factors that collectively determine overall UX. Good UX in LMS is characterised by an intuitive interface that meets a user’s psychological need for autonomy, competence and relatedness, as well as by facilitating efficient interactions (Hassenzahl 2008; International Organization for Standardization [ISO] 2010). In this context, a well-designed LMS creates a user-friendly, enjoyable learning environment that directly supports educational goals. Key elements that enhance UX include:

Alternatively, factors such as excessive complexity, lack of accessibility and system instability can frustrate users and impair the learning experience (Liu et al. 2022; Ramesh, Vermette & Chilana 2021; Vandeyar 2020). Recent research suggests that enhancing attractiveness, efficiency and dependability can alleviate these challenges (Ibrahim & Aziz 2022).

Overall, addressing these UX challenges in LMS platforms calls for the incorporation of user-centred design principles that focus on making systems more accessible, interactive and responsive through effective feedback mechanisms (Almusharraf 2024; Walraven 2024).

Educational chatbots have become essential tools for enhancing user support and engagement in learning environments (Rane 2023). Their ability to provide immediate, personalised assistance fosters interactive learning, making them increasingly valuable in LMSs (Walraven 2024). Chatbots serve various functions, from acting as digital tutors (Hobert & Berens 2023) and supplementing instructors (Chen et al. 2023), to supporting SRL (Ifelebuegu, Kulume & Cherukut 2023). Frequently asked question bots also help to address common queries efficiently.

Studies have highlighted chatbots’ effectiveness in student engagement, personalised learning and instructional support. A systematic review by Kuhail et al. (2023) has found that chatbots improve learning outcomes and student satisfaction through interactive and flexible learning methods. Additionally, they enhance motivation and curiosity by providing timely assistance (Jei, Al-Rahili & Al-Farani 2024). Their 24/7 availability ensures continuous support, particularly benefiting remote students (Al-Sharhan et al. 2020; Rocio & Wesley 2020).

Despite their potential, limitations such as limited student involvement in chatbot design can impact their effectiveness (Kuhail et al. 2023). However, their role in transforming education through personalised and interactive learning remains substantial.

Properly formulated design principles are fundamental in ensuring the creation of effective, user-friendly chatbots. These principles capture and generalise the knowledge gained from specific design instances, aiding in the design of artefacts that meet user needs and achieve desired objectives. Existing research on chatbot design principles has identified several key factors that contribute to effective and user-friendly chatbot interactions. These principles focus on enhancing transparency, engagement, adaptability, usability and ethical considerations to improve user experience and trust. The application contexts include virtual teams, enterprises and educational settings. A summary of existing chatbot design principles is provided in Table 1.

In addition to the general chatbot design principles outlined in Table 1, Jung, Lee and Park (2020) examined design principles for educational chatbots by analysing human–chatbot interactions. Their study identifies the key roles that chatbots play in education and presents specific design principles tailored to each of these roles. These include tutor chatbots, evaluator chatbots, responder chatbots, moderator chatbots and peer-learner chatbots. Beyond these role-specific principles, their study also emphasises broader design principles from a UX perspective. Educational chatbots should maintain consistency in platform components to ensure a seamless UX. It is advised that they should further incorporate feedback mechanisms to minimise user waiting time and to enhance responsiveness. A natural conversational flow, including the use of appropriate humour, can make interactions more engaging and user-friendly. Additionally, chatbots should provide clear user guidance and feature an intuitive interface to enhance accessibility and ease of use. By adhering to these principles, educational chatbots can be designed to optimise learning experiences and improve overall user satisfaction.

Integrating chatbots into LMSs presents distinct challenges that go beyond those encountered in general educational applications. The key challenges include:

While chatbot design principles provide a structured framework, LMS integration requires overcoming these distinct challenges to ensure effectiveness.

The UX Honeycomb Framework (Morville 2004) served as the conceptual framework for this study, providing a comprehensive structure to explore and evaluate user experience beyond basic usability by encompassing multiple interconnected facets that shape the overall interaction. This model functions as a qualitative tool to assess how effectively a product or service meets user needs, ensuring a holistic and satisfying experience. By emphasising the integration and balance of various UX elements, the Honeycomb Model helps identify overlooked aspects in the early design stages, making it particularly useful for refining chatbot design principles within LMS platforms.

Additionally, the literature suggests that mapping out tentative design principles in the UX honeycomb aids in systematically addressing critical UX factors (Kim 2020). The model also acts as a checklist for designers, helping them to prioritise key objectives that align with user expectations (Desmet & Hekkert 2007; Kim 2020). Furthermore, by enabling comparisons across different cases, the Honeycomb Model facilitates the development of practical guidelines that enhance user satisfaction and engagement (Lee & Kim 2017). Given these advantages, this study adopts Morville’s UX Honeycomb Model (see Figure 1) to ensure the use of a holistic approach to designing chatbots for educational environments.

The Honeycomb Model includes seven interconnected hexagons, each representing a key facet of UX. These facets include: Usefulness, Usability (how easy, effective, and efficient the product is in helping users achieve their objectives), Desirability (the appeal a product holds for the users), Findability (how easily users can find, navigate, and access the information or features), Accessibility (the extent to which a product can be used by people with diverse abilities, including those with disabilities), Credibility (the degree to which users trust the product and the information it provides), and Value (the meaningful benefits the product delivers to users and stakeholders) (Morville 2004).

Morville’s Honeycomb Model underscores the importance of harmonising these facets to create a successful and engaging UX. It serves as a qualitative measurement tool, which is especially useful in the early design stages to ensure that no aspect is overlooked (Lee & Kim 2017).

During the awareness phase, the researcher conducted a comprehensive exploration to understand the research problem, identify challenges, and determine gaps and inefficiencies in the existing landscape. This phase highlighted a significant gap: the absence of comprehensive guidelines in empirical research for designing and developing LMS chatbots. The data collection methods included personal experiences, informal discussions with other users and a literature review. Anecdotal evidence from first-hand or second-hand reports provided valuable insights into the real-life issues faced by service providers.

This phase draws on the existing literature on design principles and features. The need for these principles emerged during the awareness phase of this research, where empirical evidence highlighted gaps in LMS chatbot guidelines. To address these gaps, the proposed design principles were formulated based on the existing literature and were enriched thereby. Additionally, Morville’s (2004) UX Honeycomb Model was integrated to ensure a holistic approach to UX enhancement.

A Design Thinking workshop was initiated by the researcher, as part of the third phase of the study, conducted on 25 May 2023 at the University of Pretoria, Hatfield Campus, with 12 participants drawn from diverse South African higher education contexts. The participants included four students, four lecturers and four LMS administrators (which included the scope of instructional designers, technical support staff and curriculum developers) from both public and private institutions, as well as universities of technology and traditional universities. Using their diverse perspectives, the team created a basic prototype chatbot using Figma software. As regular LMS users, they provided critical insights into the design, functionality and user experience of an LMS-integrated chatbot, highlighting its potential to improve responsiveness, accessibility and the overall learning experience. The data collection methods included Figma screenshots, observations and field notes. The researcher conducted non-participatory observations from an adjoining room, separated by a one-way glass window, to maintain an impartial perspective and gather rich qualitative data. This approach documented the practical application of the five stages of design thinking developed by the Hasso Plattner Institute of Design at Stanford University (see Figure 3) (Meinel et al. 2022). Using their diverse perspectives, the team created a basic prototype chatbot using Figma software. It is important to note that this prototype served only as the visualisation of a potential LMS chatbot and not as the final design artefact. The design thinking workshop served as a dynamic platform for data collection, aligning with the principles of The Hasso Plattner Institute of Design at Stanford University (Meinel et al. 2022). The workshop provided a better understanding of the context and practical needs identified in the literature regarding UX in the LMS environment (Parizi et al. 2022). The data collected on tentative design principles is presented in Table 2.

During this phase, feedback on the tentative design principles was gathered using a questionnaire from four experts in LMS design. These experts were presented with a real-life LMS chatbot, specifically the Blackboard Chatbot (Anthology Inc. 2023), to illustrate the application of some of the design principles and identify any missing elements. The Blackboard Chatbot, available to Software as a Service (SaaS) clients in North America with certain features activated, provided a practical context for evaluating the principles. Permission from Blackboard was granted to allow the experts access to the system, which enhanced the real-world relevance of the evaluation. The focus of the evaluation was on refining the tentative design principles, rather than on assessing the chatbot itself. The real-life chatbot served as a tool to demonstrate the principles in action and to pinpoint areas where the principles could be better implemented or further developed.

The first section of the questionnaire presented Likert-type questions, which were directly associated with the tentative design principles outlined in Table 3. Each question evaluated the importance of the specific design principle as perceived by the expert.

Expert evaluations of the tentative LMS chatbot design principles yielded both qualitative insights and quantitative ratings that were instrumental in refining the final framework. The experts emphasised that features such as responsive interaction and focused conversation are critical for delivering timely, relevant information while effectively managing user expectations. They noted that including opt-out functionality and using an appropriate tone enhances user autonomy and engagement. Additionally, the evaluations highlighted the importance of progress tracking, personalised recommendations, and robust privacy and error-handling measures – including human-agent handover – to build trust and ensure seamless interactions.

The ‘Conclusion’ phase involved articulating the research findings as a set of design principles to be presented in an academic paper. This phase not only expands the existing knowledge base within the Information Systems discipline but also provides actionable guidance for practitioners in the field of LMS design.

This study was approved by the Research Ethics Committee of the Faculty of Engineering, Built Environment and Information Technology at the University of Pretoria (Ref. No. EBIT/246/2022). Ethical clearance was obtained in accordance with the University’s guidelines, which emphasise voluntary participation, informed consent, and the protection of participant privacy and confidentiality.

The recruitment of participants for the design thinking workshop was conducted purposively by the researcher. Participants were selected based on their active engagement with LMS and their professional roles within higher education institutions, ensuring that they possessed relevant experience and insights into the design, functionality and user experience of LMS-integrated chatbots. Recruitment was carried out directly by the researcher via email invitations, without the involvement of an independent recruiter. Prior to participation, all individuals were provided with an invitation letter outlining the study’s purpose, data collection instruments and intended use of the findings. Written informed consent was obtained from all participants before the commencement of the data collection. Participants were further informed of their right to decline participation at any stage without consequence, and that no incentives would be offered to induce participation.

To protect anonymity, the participants’ real names were concealed and replaced with pseudonyms on all transcripts and data records. Confidentiality was ensured by restricting access to identifying information to the researcher only. All collected data were securely archived at the University of Pretoria in an organised and retrievable format to ensure an audit trail and to facilitate future access, if required. The researcher maintained sensitivity and respect throughout the process to minimise any potential emotional discomfort. She also clarified that the research was conducted purely for academic purposes.